Download Malignant pleural effusion cells show aberrant glucose metabolism gene expression C-C. Lin*

Eur Respir J 2011; 37: 1453–1465
DOI: 10.1183/09031936.00015710
CopyrightßERS 2011
Malignant pleural effusion cells show
aberrant glucose metabolism gene
expression
C-C. Lin*,#, L-C. Chen", V.S. Tseng", J-J. Yan+, W-W. Lai1, W-P. Su*,#, C-H. Line,
C-Y.F. Huang** and W-C. Su*,#
ABSTRACT: Malignant pleural effusion (MPE) accompanying lung adenocarcinoma indicates
poor prognosis and early metastasis. This study aimed to identify genes related to MPE formation.
Three tissue sample cohorts, seven from healthy lungs, 18 from stage I–III lung adenocarcinoma with adjacent healthy lung tissue and 13 from lung adenocarcinomas with MPE, were
analysed by oligonucleotide microarray. The identified genes were verified by quantitative realtime PCR (qRT-PCR), immunohistochemical staining, and immunofluorescence confocal microscopy.
20 up- or down-regulated genes with a two-fold change in MPE cancer cells compared to
healthy tissues were differentially expressed from early- to late-stage lung cancer. Of 13 genes
related to cellular metabolism, aldolase A (ALDOA), sorbitol dehydrogenase (SORD), transketolase (TKT), and tuberous sclerosis 1 (TSC1) were related to glucose metabolism. qRT-PCR
validated their mRNA expressions in pleural metastatic samples. Immunohistochemical staining
confirmed aberrant TKT, ALDOA, and TSC1 expressions in tumour cells. Immunofluorescence
confirmed TKT co-localisation and co-distribution of ALDOA with thyroid transcription factor 1positive cancer cells. TKT regulated the proliferation, vascular endothelial growth factor secretion
in vitro and in vivo vascular permeability of cancer cell.
Glucose metabolic reprogramming by ALDOA, SORD, TKT and TSC1 is important in MPE
pathogenesis.
KEYWORDS: Glucose metabolism genes, lung adenocarcinoma, malignant pleural effusion
ung cancer is the leading cause of cancer
death in both males and females in the
USA, Europe and Taiwan [1]. Its incidence
is increasing and females are more likely than
males to have an adenocarcinoma subtype [2],
which is often complicated by malignant pleural
effusion (MPE) [3]. A recent study reports that
the majority of 136 patients with lung adenocarcinoma and MPE were female (61%) [4]. Patients
with MPE have been previously classified as
stage IIIB but recent data from the International
Association for the Study of Lung Cancer
propose that the staging of lung cancers with
concomitant MPE should be reclassified as early
metastasis (M1a) [5].
L
Unlike other solid cancers with surrounding
vascular structures to provide conduits for travel
and nutrient delivery, cancer cells in MPE proliferate autonomously and have a high metastatic
potential. Due to their specific biological properties,
malignant cells within MPE are uniquely capable of
surviving and proliferating without a solid-phase
scaffolding [6]. Another important mediator is the
vascular endothelial growth factor (VEGF), which
contributes to the formation of malignant effusions
by increasing vascular permeability [7]. Some genes
in MPE may be so specific in maintaining the
survival of cancer cells in MPE and VEGF secretions that MPE remains intractable and resistant to
chemotherapy [8]. A previous study reveals that
the interleukin-6/signal transducer and activator of
transcription 3/VEGF pathway [9] plays a key role
in MPE formation. A connection between EGFR
gene mutation and MPE has also been reported [4]
but the specific genes or pathways dysregulated
remain unexplored.
This study aimed to identify specific genes or
pathways of adenocarcinoma with MPE and
This article has online supplementary material available from www.erj.ersjournals.com
EUROPEAN RESPIRATORY JOURNAL
VOLUME 37 NUMBER 6
AFFILIATIONS
*Graduate Institute of Clinical
Medicine, College of Medicine,
National Cheng Kung University,
"
Dept of Computer Science and
Information Engineering, College of
Engineering, National Cheng Kung
University,
Depts of #Internal Medicine,
+
Pathology,
1
Surgery, Hospital and College of
Medicine, National Cheng Kung
University, Tainan,
e
Institute of Microbiology and
Immunology, and
**Institute of Clinical Medicine,
National Yang-Ming University,
Taipei, Taiwan.
CORRESPONDENCE
W-C. Su
Division of Hematology/Oncology
Dept of Internal Medicine
Hospital and College of Medicine
National Cheng Kung University
College
138 Sheng-Li Road
Tainan 704
Taiwan
E-mail: [email protected]
Received:
Jan 29 2010
Accepted after revision:
Sept 16 2010
First published online:
Sept 30 2010
European Respiratory Journal
Print ISSN 0903-1936
Online ISSN 1399-3003
c
1453
THORACIC ONCOLOGY
C-C. LIN ET AL.
investigate their significance in MPE cancer cell survival and
formation. It focused on female lung adenocarcinoma patients
with MPE.
MATERIALS AND METHODS
MPE cancer cells, healthy normal lung tissue, and stage
I–III lung cancer for microarray analysis
First, to identify differently expressed genes in MPE cancer cell
and normal lung tissue, two cohorts for microarray analysis
were collected. The institutional review board approved the
study and all patients provided written informed consent. The
first cohort of MPE was obtained from 13 females who
underwent thoracentesis or thoracotomy at the National
Cheng Kung Hospital (Tainan, Taiwan) from 2002 to 2005.
Cytological analysis or pathological proof from pleural biopsy
was used to verify each lung adenocarcinoma-associated MPE
specimen. The methods used for collection of cancer cells from
MPE [10], RNA extraction, Affymetrix array hybridisation, and
image processing are shown in the online supplementary
material. A second cohort of healthy lung tissue microarray
was provided by C-H. Lin (National Yang-Ming University,
Taipei, Taiwan).
Secondly, the study aimed to show that MPE-specific genes
were not only different from health lung tissue but also
relatively up- or down-regulated compared to the primary
tumour. The identified genes were MPE-specific and more
metastatic and invasive than the primary tumour. However,
the primary pulmonary tumour of MPE was often not
available and no pairing analyses (primary tumour and
MPE) could be done. Thus, a third cohort microarray database
set (National Center for Biotechnology Information gene
expression omnibus accession number: GSE7670) [11], which
included samples from 18 females with lung adenocarcinoma
with adjacent healthy tissue. The basic and clinical characteristics of these three cohorts are listed in table 1.
Statistical analysis of microarray data
Using the Golub criteria to identify differentially expressed
genes between MPE and healthy normal lung tissue [13], the
top 500 differential genes corresponding to 631 microarray
probes were identified. These 500 discriminative genes (MPEspecific genes) were further analysed using the third cohort
microarray database set. The first and the third cohorts were
pooled and classified as a fourth group including samples
from 18 subjects with adjacent normal lung tissue, nine with
TABLE 1
Characteristics of the three female cohorts for
array analysis
Variable
Stage#
Age yrs
Normal healthy females
18 lung cancer
58 (48–72)
9 stage I (3 IA and 6 IB)
60 (48–73)
9 stage III (5 IIIA and 4 IIIB)
13 MPE
5 IIIB
58 (43–79)
8 IV
Age values are given as mean (range). MPE: malignant pleural effusion.
#
: stage according to American Joint Committee on Cancer [12].
1454
VOLUME 37 NUMBER 6
stage I lung cancer, nine with stage III, and 13 with MPE cancer
cells, representing normal tissue to early stage lung cancer to
late-stage (or pre-metastatic) stage.
After using robust multi-array analysis normalisation [11], the
631 probes were further restricted by adding the criterion that
the genes show a tendency to be expressed at higher or lower
levels between the adjacent healthy, stage I, stage III and MPE
samples. First, the identified genes had to be differentially
expressed between the four groups. One-way ANOVA was
used to search for genes that were expressed statistically
differently among these four groups (p,0.05). Moreover, the
expressions of these genes had to be higher in MPE compared
with the tumour part of stage III cancer, higher in the tumour
part of stage III cancer compared with that of stage I cancer,
and higher in the tumour part of stage I cancer than the
adjacent normal lung tissue. Otherwise, the expression of these
genes had to be lower in MPE compared with the tumour part
of stage III cancer, lower in the tumour part of stage III cancer
compared with that of stage I cancer, and lower in the tumour
part of stage I cancer than the adjacent normal lung tissue.
The identified up- or down-regulated genes in MPE cancer
cells were also expressed differentially as the cancer progressed from early stage to MPE. Using these supervised and
strict set of criteria among the four groups, 71 genes were
identified from 500 genes. These 71 genes were further
restricted to the expression level at least two-fold higher in
MPE samples than normal lung tissue, with p,0.001 to
exclude false significant results [14]. The workflow for the
analysis is shown in figure 1a.
RNA extraction and quantitative RT-PCR
RNA extraction and quantitative RT-PCR were done by Welgene
Biotech Co., Ltd (Taipei, Taiwan), while total RNA purification
was done by TRIzol reagent combined spin columns. All samples were PCR-amplified (LightCycler-FastStart DNA Master
SYBR Green I kit; Roche Diagnostics Corporation, Roche
Applied Science, Indianapolis, IN, USA) with constitutively
expressed b-actin. The primers designed for the identified genes
were listed in table 2.
Immunohistochemistry and immunofluorescent images
Immunohistochemistry and immunofluorescence analysis validated the gene expression in clinical specimens. The basic and
clinical characteristics of the three validation cohorts were
listed in table 3. The MPE cell blocks were produced using the
AgarCyto cell block method [15]. The methods and the
antibodies used for immunohistochemistry and immunofluorescence are shown in the online supplementary material.
Cell lines, cell lysates and Western blot
The PC14PE6/AS2 (AS2) cell line was from ascites generated
from intraperitoneal injection with PC14PE6 in the metastatic
animal model [9]. The normal bronchial cell line (NL-20), lung
adenocarcinoma cell line A549, and H1650 were purchased
from American Type Culture Collection (Rockville, MD, USA).
The CL1-0 clonal cell line was provided by P-C. Yang (National
Taiwan University College of Medicine, Taipei, Taiwan).
Total protein from healthy lung lysates and cancer cell lines
were extracted, blotted and detected using antibodies against
EUROPEAN RESPIRATORY JOURNAL
C-C. LIN ET AL.
THORACIC ONCOLOGY
a)
7 females
Normal lung
13 females
MPE
Two-group comparative analysis
Top 500 by Golub method
631 probe
18 adjacent
Normal chip
9 stage I
(3 IA and 6 IB)
9 stage III
(5 IIIA and 4 IIIB)
13
MPE
ANOVA analysis and post hoc multiple
comparisons–Bonferroni method
MPE > stage III > stage I > normal
or MPE < stage III < stage I < normal
71 genes
MPE versus normal
p<0.001
Relative change >2-fold
20 gene
b)
Adjacent normal
Stage 1
Stage 3
MPE
TCF7L1
PANK2
ECHDC2
TSC1
CALCOCO1
TSC22D1
C3orf24
MAPK13
MRPL15
NME1
MTX2
SORD
C20orf24
P4HA1
ALDOA
POLE3
SLBP
BZW1
CNIH
TKT
0.25
1.0
4.0
Relative expression to median
FIGURE 1.
a) Flow chart for the collection of three different cohorts for microarray analysis. b) Gene expression profiles clustered hierarchically based on final 20 genes.
MPE: malignant pleural effusion.
EUROPEAN RESPIRATORY JOURNAL
c
VOLUME 37 NUMBER 6
1455
THORACIC ONCOLOGY
TABLE 2
C-C. LIN ET AL.
RT-PCR primers of selected genes for the
Lightcycler
Gene symbol
Sense primer
Antisense primer
TSC1
TGGGAATTGGAATCAAAAGAG ACAAGCAACTGCCTTGACATT
SORD
TGACCACCGTACCCCTACTG
ALDOA
GGCCTCCGTCTGGATTTC
GGGCATGGTGCTGGTAGTAG
TKT
ATGCCATTGCACAAGCTG
CACACTTCATACCCGCCCTA
PANK2
GGTCTTGGCAATCATCTGTG
CCCTTCCAAAAACTGCTTGT
CAGACTTGGACGCAAGCAT
human transketolase (TKT), aldolase A (ALDOA; Proteintech
Group, Chicago, IL, USA), tuberous sclerosis 1 (TSC1; Cell Signaling Technology, Inc., Beverly, MA, USA), and glyceraldehyde-3phosphate dehydrogenase (Santa Cruz Biotechnology, Inc, Santa
Cruz, CA, USA). Antibody binding was detected by electrochemiluminescence (Amersham Biosciences, Piscataway, NJ,
USA) based on the manufacturer’s instructions.
TKT inhibitor and MTT test
Oxythiamine was purchased from Sigma Chemical Co. (St
Louis, MO, USA). A 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) test was used to evaluate the antiproliferation effect of oxythiamine.
Transfection with siRNA, proliferation analysis, colony
formation assay, in vitro VEGF secretion and in vivo
vascular permeability assay
Oligonucleotides representing small interfering (si)RNA against
human TKT expression (TKT siRNA) and mismatch control
oligonucleotides (scramble siRNA) were used (Invitrogen-Life
Technologies, Carlsbad, CA, USA). The siRNA sequence that
targeted TKT was 59-AAAGAGGACAGCCAUGAUCUCUG
CG-39, while the scramble siRNA was the negative control.
The PC14PE6/AS2 cells were transfected with siRNA for a final
concentration of 50–100 nM by MicroPorator MP-100 (Nano
EnTek, Seoul, South Korea).
The cell proliferation of PC14PE6/AS2 after transfection was
detected by flow cytometry analysis with proliferation-associated antigen Ki-67 and cell colony formation assay (online
supplementary material).
TABLE 3
Transfected PC14PE6/AS2 cells were maintained in 60 mm
ultra-low attachment plate (50,000 cells per plate; Corning,
Lowell, MA, USA) for 48 h to evaluate cell morphology, VEGF
secretion in vitro and in vivo permeability assay (Miles
permeability assay; online supplementary material) [9].
Statistical analysis
Data were analysed using Prism 4 (GraphPad Software, Inc.,
La Jolla, CA, USA). Differential expressions of specific genes
between different stages were assessed by one-way analysis of
variance (ANOVA) and by Bonferroni post hoc multiple
comparisons. The area grade of immunohistochemistry
between tumour and healthy samples was determined by the
Mann–Whitney U-test. Paired t-test was used to determine
differences between TKT siRNA and scramble-transfected
PC14PE6/AS2 cells. Statistical significance was set at p,0.05.
RESULTS
Glucose metabolism regulatory genes were aberrantly
regulated in MPE cancer cells
20 differentially expressed genes were identified between MPE
and healthy adjacent lung tissue. 14 genes were up-regulated
and six were down-regulated in the MPE group compared
with the healthy controls (fig. 1b). Their relative changes and
annotation are listed in table 4. Functional annotation and
pathway interaction of these genes was done using the
BABELOMICS platform [16] and ingenuity pathways analysis
(IPA) [17]. Gene ontology annotation of three of the 20 genes
remained unestablished.
Combining IPA canonical pathway analysis and functional
definitions with gene ontology and KEGG pathway analysis
in the BABELOMICS platform, 13 of the 17 well-known genes
(76%) were related to metabolic processes. Further studies of
the 13 genes showed that three, ALDOA, sorbitol dehydrogenase (SORD) and TKT, were directly related to glucose
metabolism (fig. 2a and b). Of the down-regulated genes,
TSC1 was involved in the insulin pathways, according to the
KEGG database.
Whether these genes are differentially expressed from healthy
to early stage lung cancer to MPE were further investigated.
The difference between stage I cancer and MPE was
statistically significant (fig. 2c; p,0.01 in SORD and p,0.05
in ALDOA and TKT, all by Bonferroni multiple comparisons).
Characteristics of three cohorts for validation
Variable
Immunohistochemistry of normal lung tissue and adjacent
Sex
Stage
Age yrs
28 males
11 stage IA, 8 stage IB
57 (30–92)
20 females
7 stage IIA, 6 stage IIA
6 stage IIIA, 2 stage IIIB
8 stage IV
10 metastatic pleurae
8 MPE block
6 females
4 stage IIIB
4 males
6 stage IV
4 females
1 stage IIIB
4 males
7 stage IV
57 (26–85)
63 (46–75)
Age values are given as mean (range). MPE: malignant pleural effusion.
1456
VOLUME 37 NUMBER 6
EUROPEAN RESPIRATORY JOURNAL
C-C. LIN ET AL.
TABLE 4
Clone
THORACIC ONCOLOGY
List of differentially expressed genes between malignant pleural effusion and normal lung tissue#
Symbol
Description
Relative change
4.48
201563_at
SORD
Sorbitol dehydrogenase
200776_s_at
BZW1
Basic leucine zipper and W2 domains 1
4.11
208699_x_at
TKT
Transketolase
3.38
201577_at
NME1
Non-metastatic cells 1, protein (NM23A)
3.29
219288_at
C3orf14
Chromosome 3 open reading frame 14
3.25
200966_x_at
ALDOA
Aldolase A
3.14
218027_at
MRPL15
Mitochondrial ribosomal protein L15
2.78
210058_at
MAPK13
Mitogen-activated protein kinase 13
2.61
206052_s_at
SLBP
Stem-loop (histone) binding protein
2.58
217835_x_at
C20orf24
Chromosome 20 open reading frame 24
2.48
207543_s_at
P4HA1
Procollagen-proline, 2-oxoglutarate 4-dioxygenase (proline 4-hydroxylase)
2.42
208828_at
POLE3
Polymerase (DNA directed), epsilon 3 (p17 subunit)
2.31
201653_at
CNIH
Cornichon homologue
2.19
203517_at
MTX2
Metaxin 2
2.02
TCF7L1
Transcription factor 7-like 1
0.47
TSC1
Tuberous sclerosis 1
0.45
CALCOCO1
Calcium binding and coiled-coil domain 1
0.44
ECHDC2
Enoyl coenzyme A hydratase domain containing 2
0.40
PANK2
Pantothenate kinase 2
0.36
TSC22D1
TSC22 domain family, member 1
0.23
221016_s_at
209390_at
209002_s_at
218552_at
218809_at
215111_s_at
#
: relative change .2- or ,0.5-fold; p,0.001.
The expression levels were also significantly different between
the healthy and stage III samples in the three genes (p,0.01 in
ALDOA, and p,0.05 in SORD and TKT). Expression levels of
TSC1 were lower in the MPE samples than in the stage I
(p,0.01) and III (p,0.05) samples.
To verify these gene expression data, RNA was extracted from
nine metastatic pleurae lung adenocarcinoma and adjacent
normal lung tissue. Samples were then analysed using
quantitative RT-PCR. ALDOA and TKT expressions were
significantly higher, while TSC1 and pantothenate kinase 2
(PANK2; a non-glucose-metabolism gene) expressions were
significantly lower in tumour than in healthy samples (fig. 2d;
p,0.05 by t-test). SORD expression was also higher in tumour
samples, but the difference was not statistically significant
(p50.05).
TKT and ALDOA protein expression was higher and TSC1
protein expression was lower in tumour cells than in
healthy cells
Western blotting was used to test the protein expression levels
of TKT, ALDOA and TSC1 in four lung cancer cell lines: A549,
PC14PE6/AS2, H1650 and CL1-0. TKT and ALDOA expression
were higher and TSC1 expression was lower in lung cancer cell
lines than in healthy lung lysates (fig. 2e) and normal bronchial
cell line (NL-20; online supplementary material). Immunohistochemical staining validated the protein expression in
clinical samples of lung cancer tissue, adjacent healthy tissue,
tumours invading the pericardium, and cancer cells from
patients with MPE. TKT was located predominantly in the
nucleus (fig. 3a and b) and was more abundantly expressed in
tumour cells that invaded the pericardium (fig. 3c), and
tumour tissue than in adjacent healthy bronchial epithelium.
EUROPEAN RESPIRATORY JOURNAL
Cancer cells from MPE were used to make AgarCyto cell
blocks and to study TKT expression in MPE cancer cells.
Cancer cells in MPE expressed high levels of TKT (fig. 3d).
ALDOA expression, located predominately in cytoplasm, was
higher in tumour cells than in adjacent healthy bronchial
epithelial cells (fig. 3e and f). Tumour cells that invaded the
pericardium also expressed high ALDOA levels (fig. 3g).
Conversely, TSC1, located primarily in the cytoplasm and
nuclei, was expressed mainly in healthy pneumocytes, but
barely or not at all in tumour cells (fig. 3h and i).
Immunohistochemical analyses of 48 lung adenocarcinoma
specimens showed that 19 were stage I, 13 stage II, eight stage
III and eight stage IV. The percentage of immunoreactive
cancer cells was scaled using a four-tiered area grade and
semi-quantitative system [18]. Using the area grade, the
staining scores for TKT (fig. 3j; p,0.001) and ALDOA
(p,0.01) were significantly higher in tumour tissue than in
adjacent healthy tissue, and the staining score for TSC1 was
significantly lower in tumour tissue (p,0.001) than in adjacent
healthy tissue. A limited number of samples for each stage
prevented further correlation studies of staining scores and
cancer stages.
Cancer cell in MPE and pleural metastasis expressed high
TKT and ALDOA
Thyroid transcription factor 1 (TTF-1) served as a good marker
of lung cancer cells in MPE [19, 20]. To verify TKT and ALDOA
expressions in lung cancer cells from MPE, eight MPE blocks
from eight patients who underwent thoracocentesis and cancer
cells were examined by cytology (table 2). Triple immunostaining using DAPI, TTF-1, TKT or ALDOA and recording
under confocal microscope were used to investigate the
VOLUME 37 NUMBER 6
1457
c
THORACIC ONCOLOGY
C-C. LIN ET AL.
a)
Biological process; Level 3
20
0
40
KEGG pathway
0
80
100
76.47% Fructose and mannose metabolism
Pyrimidine metabolism
70.59%
60
Cellular metabolic process
Macromolecule metabolic process
Primary metabolic process
70.59%
20
0.50
0.75
2.50
2.75
Carbon fixation
Pentose phosphate pathway
10.53%
-log(p-value)
1.25 1.50 1.75
1.00
80
100
10.53%
10.53%
Canonical pathway analysis
0.25
60
Purine metabolism
10.53%
b)
0.00
40
10.53%
2.00
2.25
3.00
Pentose phosphate pathway
Fructose and mannose metabolism
Pyrimidine metabolism
Parkinson’s signalling
Pantothenate and CoA biosynthesis
Purine metabolism
Threshold
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11 0.12 0.13 0.14 0.15
Ratio
c)
SORD
ALDOA
Normalised
expression value
3
1
2
2
1
1
0
Normal Stage 1 Stage 3 MPE
2
3
2
0
1
0
0
Normal Stage 1 Stage 3 MPE
Normal Stage 1 Stage 3 MPE
e)
4
Lung lysates AS2
Normal Stage 1 Stage 3 MPE
A549
H1650
CL1-0
AS2
A549
CL1-0
TSC1
***
log2 change (relative)
TSC1
4
3
d)
TKT
TKT
SORD
**
2
ALDOA
ALDOA
TKT
GAPDH
PANK2
0
Lung lysates
**
***
TSC1
GAPDH
-2
FIGURE 2.
Analysis of all genes and selected gene expression in different stages of lung cancer via bioinformatics platform, qRT-PCR of tissue samples and Western
blotting of lung cancer cell lines. a) The BABELOMICS platform revealed that most genes were associated with cell metabolism, according to the gene ontology database.
The axis indicates the percentage of well-known genes associated with a specific biological function or pathway. b) Ingenuity pathways analysis of all genes was displayed by
canonical pathway. The axis indicates the significance (-log p-value; threshold p,0.05) of the pathway association. c) Gene expression values of glycolysis-associated genes
(tuberous sclerosis 1 (TSC1), sorbitol dehydrogenase (SORD), aldolase A (ALDOA) and transketolase (TKT)) from healthy controls and lung cancer patients at different stages
(using robust multi-array analysis normalisation). *: p,0.05; **: p,0.01; ***: p,0.001. d) Quantitative real-time PCR analysis of TSC1, SORD, ALDOA, TKT and
pantothenate kinase 2 (PANK2) from four lung cancer and adjacent healthy lung tissue samples. e) Western blotting was used to detect TKT, ALDOA (PC14PE6/AS2, A549,
H1650 and CL1-0) and TSC1 (PC14PE6/AS2, A549, CL1-0) expressions in the lysates from healthy lung and lung cancer cell lines. GAPDH: glyceraldehyde phosphate
dehydrogenase.
1458
VOLUME 37 NUMBER 6
EUROPEAN RESPIRATORY JOURNAL
C-C. LIN ET AL.
THORACIC ONCOLOGY
a)
b)
c)
d)
e)
f)
g)
h)
i)
TKT
j)
4
ALDOA
1.5
Tumour
TSC1
1.5
Tumour
Normal
Normal
Tumour
Normal
2
1.0
Area grade
Area grade
Area grade
3
0.5
1.0
0.5
1
0
FIGURE 3.
0.0
0.0
Immunohistochemical staining of transketolase (TKT), aldolase A (ALDOA) and tuberous sclerosis 1 (TSC1) in clinical specimens. a) TKT intensity in healthy
bronchiole epithelium (arrows, 206 objective). b) Nucleus immunoreactive TKT predominantly expressed in the lung adenocarcinoma adjacent to healthy bronchiole
epithelium (arrowheads, 206 objective). c) TKT expression in lung cancer cells invading the pericardium (406 objective). d) TKT intensity of cancer cells in malignant pleural
effusion (406 objective). e) Cytoplasmic localisation of predominant ALDOA in lung adenocarcinoma (arrowhead, 206 objective) adjacent to the healthy bronchiole
epithelium (arrows, 206 objective). f) ALDOA intensity in lung adenocarcinoma. g) ALDOA expression in lung cancer cells invading the pericardium (406 objective). h)
Nuclear and cytoplasmic expression of TSC1 in pneumocytes (arrowheads, 206 objective). i) No expression of TSC1 in adjacent tumour tissue (206 objective). Scale
bars510 mm. j) Immunohistochemical analysis of TKT, ALDOA, and TSC1 expression in tumour and adjacent healthy tissue specimens from 48 patients (semi-quantitative
area scores: 0, no stained tumour cells; 1, 0–25% of tumour cells stained; 2, 26–50% tumour cells stained; 3, 51–75% tumour cells stained; 4, 76–100% tumour cells stained).
**: p,0.01; ***: p,0.001 (paired t-test). Higher TKT, higher ALDOA, and lower TSC1 expression was observed in tumour tissue compared to the adjacent healthy area.
EUROPEAN RESPIRATORY JOURNAL
VOLUME 37 NUMBER 6
1459
c
THORACIC ONCOLOGY
C-C. LIN ET AL.
expression and distribution of TKT and ALDOA in the TTF-1positive cells from MPE blocks. TTF-1 expression was
predominant in the nucleus of cancer cells from MPE block
while TKT immunoreactivity was detected in the nuclei
located within TTF-1-positive cancer cells (fig. 4a).
The continuous section showed that ALDOA expression was
detected in the cytoplasm and cell membrane within TTF-1positive cells (fig. 4b). Co-localisation of TKT and co-expression
of ALDOA in TTF-1-positive cancer cells were also observed
under lower power field with the acinar type (fig. 4c). Further
quantitative analysis of ALDOA and TKT expression in the
eight MPE cell blocks showed that the co-localisation ratios of
TKT and TTF1 were .70% (fig. 4d; using pixel-by-pixel
analysis through by FV-1000 software) [21]. Defining coexpression of ALDOA as the distribution of ALDOA staining
in more than one-third of the plasma membrane of TKT positive
cells [21], the co-expression ratios were .75% (fig. 4d).
There were high TTF-1, TKT and ALDOA expressions in
metastatic pleural tissue by immunohistochemistry staining
(fig. 5a). Immunofluorescence staining and photography by
confocal microscope on the continuous section further confirmed the co-localisation of TKT and co-expression of ALDOA
in TTF-1-positive cancer cells (fig. 5b). Immunohistochemical
analyses of the metastatic pleural adenocarcinoma of 10
patients (table 4) showed that staining scores for TKT and
ALDOA were significantly higher in tumour tissue compared
to adjacent stroma and endothelial cell (fig. 5c and online
supplementary material; p,0.001).
Oxythiamine and TKT knockdown with siRNA inhibited lung
cancer cell proliferation
Oxythiamine, a well-known TKT inhibitor, blocked TKT
function by competing with thiamine, a key TKT co-factor
[22]. It dose-dependently inhibited the proliferation of AS2,
A549, CL1-0 and H1650 lung cancer cells 96 h after being
added to the culture medium (fig. 6a). MTT assays showed
that the half maximal inhibitory concentrations of all cells were
around 10–20 mM.
To verify the role of TKT on cancer cell proliferation, siRNA
was used to knockdown TKT expression in PC14PE6/AS2
cells. The transfection of TKT siRNA into cells suppressed TKT
expression but not in the scrambled and untransfected control,
as assayed by Western blotting (fig. 6b). To investigate the
effects of proliferation, the proportion of ki-67 staining (fig. 6c)
was lower in TKT siRNA transfected cells (54%) than in the
scramble control (74%). The rate of colony growth also
decreased (20%) by colony formation assay (fig. 6d).
TKT knockdown with siRNA prevented lung cancer cell
VEGF secretion in vitro and vascular permeability in vivo
Since cancer cells in the pleural cavity tend to be spheroid or
clustered, but are flattened attachments when cultured in
adherent culture dishes, PC14PE6/AS2 cells were maintained
in ultra-low attachment plates to investigate the effects on
knockdown TKT. Using TKT siRNA to knockdown TKT
expression, TKT siRNA cells dispersed instead of forming
clusters (data not shown).
VEGF is a vascular permeability factor contributing to the
generation of malignant effusions [7]. First, ELISA of VEGF
1460
VOLUME 37 NUMBER 6
secretion of the conditioned medium revealed decreased VEGF
secretions in si-TKT PC14PE6/AS2 cells compared to the
si-scramble controls (fig. 7a).
The in vivo permeability assay (Miles permeability assay) was
performed to determine if knockdown of TKT can downregulate VEGF and lead to diminished vessel permeability.
The areas of dye leakage induced by conditioned medium
from TKT siRNA PC14PE6/AS2 cells were smaller than the
leakage area from scramble transfected cells (fig. 7b and c).
Thus, VEGF produced by PC14PE6/AS2 cells is biologically
active, and that down-regulation of VEGF secretion in
PC14PE6/AS2 cells by inhibiting TKT can reduce vascular
permeability.
DISCUSSION
20 genes are not only up-regulated (14) or down-regulated (six)
in MPE cancer cells compared to normal lung tissue but also
expressed differentially as lung cancer progresses from early
stage to MPE. Of the 13 genes related to cell metabolism, three
(ALDOA, SORD and TKT) are directly involved in glucose
metabolism while one (TSC1) is involved in the insulin
pathway. Unlike other microarray studies derived directly
from solid tumours, the MPE in this study has been collected
from pleural metastatic lung adenocarcinoma and cancer cells
have been further enriched as in a previous study [10].
Immunohistochemistry and immunofluorescence analysis
of MPE and metastatic pleural lesion further confirm the
specificity of these genes in cancer cells other than in adjacent
fibroblast and endothelium cell.
In 1924, WARBURG [23] proposed that tumour cells use
glycolysis even in oxygen-rich environments and concluded
that this defective metabolism results in cancer formation.
Cancer cells in MPE proliferate autonomously and have a high
metastatic potential [6]. Other investigators [24] have also
demonstrated that low glucose concentrations and pH in the
pleural effusion predicts shorter survival and less successful
pleurodesis. An 18F-2-deoxyglucose (FDG) positron emission
tomography study [25] shows that higher 18F-FDG uptake
(higher glycolysis) in pleural effusion is associated with poor
survival. Other glucose degradation pathways such as the
pentose phosphate and sorbitol pathways have also been
associated with cancer proliferation [26, 27].
Recently, many basic and clinical studies have confirmed that
metabolic reprogramming benefits cancer cells [28]. Inhibiting
glycolysis-associated enzymes blocks tumour growth in animal
models. Conversely, VEGF acts as a key mediator in pleural
effusion formation by increasing vascular permeability [7].
Many studies demonstrate the link between glycolysis and
VEGF activation. Though hypoxia is a definite VEGF trigger
[29], normoxic cells or cancer cells exposed to inflammatory
stimuli, such as interferon, tumour necrosis factor-a and
interleukin, may manifest phenotypic changes like those
observed in hypoxic cells [30]. These inflammatory stimuli
also induce VEGF secretion in lung cancer with MPE [9, 31].
The product of glycolysis, lactate, has also been proven to
incite angiogenesis, even in a normoxic environment [32].
In the current study, ALDOA, TKT and SORD are all involved
in glucose metabolism. There are three aldolase isozymes in
humans and ALDOA expression is higher in lung tumours
EUROPEAN RESPIRATORY JOURNAL
C-C. LIN ET AL.
THORACIC ONCOLOGY
a)
DAPI
TTF-1
b)
DAPI
TTF-1
TTF-1+DIC
ALDOA
TKT
TTF-1+TKT
TTF-1+ALDOA
TTF-1
TKT
TTF-1+TKT
DAPI
TTF-1
ALDOA
TTF-1+ALDOA
d) 100
e)
TTF-1+ cell with ALDOA cell
membrane distribution %
DAPI
Co-localisation of TKT with
TTF-1 %
c)
DAPI+TTF-1
75
50
25
100
80
60
40
20
FIGURE 4.
MPE8
MPE7
MPE6
MPE5
MPE4
MPE3
MPE2
MPE1
MPE8
MPE7
MPE6
MPE5
MPE4
MPE3
MPE2
MPE1
0
Immunofluorescent analysis detected the expression and distribution of transketolase (TKT), aldolase A (ALDOA) and thyroid transcription factor 1 (TTF-1) at
malignant pleural effusion (MPE) lung cancer cells. a) Confocal image under higher power field (606 objective ) showed the immuno-detection of TTF-1 at nucleus
demonstrated by DAPI stain for nucleus, TTF-1 stain, merge of DAPI stain and TTF-1 stain and merge of TTF-1 and DIC (differential interference contrast) images. The colocalisation of TTF-1 and TKT was verified by TKT stain and merge of TTF-1 and TKT stain. b) Continuous section showed simultaneous immunodetection of TTF-1 and
ALDOA (cytoplasma and cell membrane) by DAPI stain, TTF-1 stain, ALDOA stain and merge of TTF-1 and ALDOA stain. c) Representative confocal images under low power
field (206 objective) showed co-localisation of TKT and TTF-1 in nucleus and cytoplasm distribution of ALDOA in TTF-1-positive MPE cancer cells. Scale bars510 mm.
d) Immunofluorescent analysis was used to detect the co-localisation ratio between TTF-1 and TKT (pixel-by-pixel analyses), and ratio of cytoplasm co-distribution of ALDOA
in TTF-1-positive cancer cell from eight patients’ MPE cell blocks.
EUROPEAN RESPIRATORY JOURNAL
VOLUME 37 NUMBER 6
1461
c
THORACIC ONCOLOGY
HE stain
a)
b)
c)
C-C. LIN ET AL.
TTF-1
TKT
DAPI
TTF-1
TKT
TTF-1+TKT
DAPI
TTF-1
ALDOA
TTF-1+ALDOA
TKT
4
ALDOA
4
Tumour
3
Area grade
Area grade
Endothelial cell
2
Tumour
Fibroblast
Fibroblast
3
ALDOA
Endothelial cell
2
1
1
0
0
FIGURE 5.
Distribution of transketolase (TKT), aldolase A (ALDOA) and thyroid transcription factor 1 (TTF-1) at metastatic pleurae tumour. a) Immunohistochemical
analysis was used to detect the expression of TKT, ALDOA and TTF-1 in metastatic pleurae tumour (area of dashed line, 206 objective) compared to adjacent normal pleurae
tissue. b) Continuous section for immunofluorescence analysis and confocal image showed co-localisation of TKT and TTF-1 in nucleus and cytoplasmic distribution of
ALDOA in TTF-1-positive cancer cells in metastatic pleurae tumour (area of dashed line, 206 objective). Scale bars520 mm. c) Immunohistochemical analysis was used to
detect the expression of TKT and ALDOA in the metastatic pleurae tumour of 10 patients. Higher expression of TKT and ALDOA was found in tumour tissue compared with the
adjacent fibroblast and endothelial cell (semi-quantitative area scores: 0, no stained tumour cells; 1, 0–25% of tumour cells stained; 2, 26–50% tumour cells stained; 3, 51–
75% tumour cells stained; 4, 76–100% tumour cells stained). ***: p,0.001.
1462
VOLUME 37 NUMBER 6
EUROPEAN RESPIRATORY JOURNAL
C-C. LIN ET AL.
a) 100
THORACIC ONCOLOGY
AS2
A549
CL1-0
H1650
80
60
40
20
0
2.5
5
10
20
Oxythiamine mM
Control
b)
40
Mock
2.5
5
10
20
Oxythiamine mM
40
2.5
5
10
20
Oxythiamine mM
40
2.5
5
10
20
Oxythiamine mM
40
Scramble siRNA-TKT
TKT
Actin
c)
0.0%
104
0.1%
0.0%
73.7%
IgG control
Ki67(+)% 55.2%
0.1%
Scramble
siRNA-TKT
103
FL1-H
Ki67(+)
102
101
Ki67(-)
100
0
0.2%
FSC-H
1023
99.7%
0
0.0%
1023
26.3%
FSC-H
0
0.5%
FSC-H
1023
44.3%
Forward scatter
d)
Colony % of control
Ki-67(+) cell %
80
AS2-siRNA-control
60
AS2-siRNA-TKT
40
100 100
AS2-siRNA-control
AS2-siRNA-TKT
80 80
60 60
40 40
20
20 20
0
FIGURE 6.
0
0
Inhibition of transketolase (TKT) blocked the viability or proliferation of lung cancer cells. a) Oxythiamine, a TKT inhibitor, blocked the viability of lung cancer
cells. The viability rate was detected using MTT test. b) The knockdown of TKT with TKT siRNA inhibited the proliferation of PC14PE6/AS2 cells. Western blotting analyses
showed the expression of TKT of PC14PE6/AS2 cells transfected or untransfected with siRNA-scramble control, siRNA-TKT. c) Histograms showed Ki-67 staining of TKT
siRNA and scramble-transfected PC14PE6/AS2 cells compared to the isotype control. Ig: immunoglobulin. d) Proportions of Ki-67 positive and colony formation in TKT siRNA
or scramble-transfected PC14PE6/AS2 cells. **: p,0.01 (paired t-test).
EUROPEAN RESPIRATORY JOURNAL
VOLUME 37 NUMBER 6
1463
c
THORACIC ONCOLOGY
C-C. LIN ET AL.
a)
AS2-siRNA-control
VEGF % of scramble control
100
AS2-siRNA-TKT
75
50
25
0
b)
siRNA-scramble
c)
15
siRNA-TKT
AS2-siRNA-control
Relative leakage area
AS2-siRNA-TKT
10
TKT expression is not only higher in MPE in microarray analysis
but also verified in primary lung cancer tissue, metastatic
pleural tumour, and malignant pleural effusion. Using the TKT
inhibitor, oxythiamine, or transfecting lung cancer cell with
small interfering RNA against TKT, inhibiting TKT suppresses
lung cancer cell proliferation. Therefore, up-regulating TKT in
cancer cells from MPE may facilitate their survival by activating
the pentose phosphate pathway. Using an ultra-low attachment
plate, the knockdown of TKT prevents cell aggregation, which is
related to colony formation in soft agar [35]. TKT also controls
the secretions of VEGF and vascular permeability in vitro and in
vivo. A recent study in a diabetic animal model also verifies that
the TKT activator benfotiamine stimulates the activity of
pentose phosphate pathway enzymes, leading to phosphorylation/activation of VEGF receptor-2 [36]. Taken together, TKT is
involved not only in MPE cancer cell proliferation but also in
VEGF secretion in MPE formation.
The TSC1/TSC2 complex is associated with the development of
tuberous sclerosis, and mutations in either gene are responsible
for both the familial and sporadic forms. Although not directly
involved in glucose metabolism, TSC1 modulates the mammalian target of rapamycin (mTOR) pathway and receives input
from the PI3K-Akt and LKB1-AMPK pathways. Hyperactivation
of mTOR alone is sufficient to drive hypoxia-inducible factordependent transcription of glycolytic and angiogenesis-related
genes [37]. The LKB1 gene is inactivated in lung cancer cells and
related to tumourigenesis and metastasis [38]. Moreover, TSC1
regulates VEGF expression via the mTOR pathway [39]. Considering all of these, the AMPK-LKB1-TSC1 pathway is a potential target for lung cancer therapy.
In conclusion, a panel of genes from MPE may contribute to
lung cancer progression. Glucose metabolism is important not
only in tumourigenesis but also in cancer cell proliferation and
VEGF secretion in MPE. This study increases the understanding of underlying mechanisms of MPE, which may facilitate
the development of new treatment strategies for MPE-associated
lung adenocarcinoma.
5
0
SUPPORT STATEMENT
FIGURE 7.
Inhibition of transketolase (TKT) in PC14PE6/AS2 cells reduced
vascular endothelial growth factor (VEGF) secretion and vascular permeability.
a) Secretion of VEGF of PC14PE6/AS2 was calculated after transfection with
This work was supported by the Department of Health (DOH99-TD-C111-003, DOH99-TD-B-111-102), and by grants NSC 97-2314-B-006-014MY2 and NHRI93A1-NSCBS04-2-5.
scramble siRNA or TKT siRNA and maintained in ultra-low attachment plate (n54).
**: p,0.01. b) Biological activity of VEGF secreted by PC14PE6/AS2 cells after
STATEMENT OF INTEREST
transfection with scramble siRNA or TKT siRNA was measured by the Miles
None declared.
permeability assay. The dye leakage areas are marked by dashed red circles.
c) The areas of dye leakage were calculated for each injection site (n54). *: p,0.05.
than in healthy tissue [33]. The mechanism for this upregulation remains unknown. SORD is involved in the sorbitol
pathway of glucose metabolism and although no studies have
linked SORD to tumourigenesis, blocking the sorbitol pathway
suppresses colon cancer cell proliferation [27]. TKT, the key
enzyme in the pentose phosphate pathway, which mediates
the conversion of glucose to ribose phosphate, is used in the
biosynthesis of nucleic acids and nucleotides. The pentose
phosphate pathway is up-regulated in cells with anchorageindependent cell growth phenotype [34], which have high
metastatic potential. Cancer cells that proliferate in MPE are
also anchorage independent.
1464
VOLUME 37 NUMBER 6
ACKNOWLEDGEMENTS
The authors thank I. J. Fidler (MD Anderson Cancer Center, Houston, TX,
USA) for the human lung adenocarcinoma cell line PC14PE6, and P-C.
Yang (National Taiwan University College of Medicine, Taipei, Taiwan)
for the human lung adenocarcinoma cell line CL1-0. The authors thank
the Microarray and Gene Expression Analysis Core Facility of the
National Yang-Ming University VGH Genome Research Center (Taipei)
for technical support and E.Y. Chuang (Bioinformatics and Biostatistics
Core, National Taiwan University, Taipei) for help with the ingenuity
pathway analysis. The authors also thank G. Alzona Nisperos for editing.
REFERENCES
1 Spiro SG, Silvestri GA. One hundred years of lung cancer. Am J
Respir Crit Care Med 2005; 172: 523–529.
EUROPEAN RESPIRATORY JOURNAL
C-C. LIN ET AL.
2 Coscio AM, Garst J. Lung cancer in women. Curr Oncol Rep 2006; 8:
248–251.
3 Trompezinski S, Denis A, Schmitt D, et al. IL-10 is unable to
downregulate VEGF expression in human activated keratinocytes.
Arch Dermatol Res 2002; 294: 377–379.
4 Wu SG, Gow CH, Yu CJ, et al. Frequent EGFR mutations in
malignant pleural effusion of lung adenocarcinoma. Eur Respir J
2008; 32: 924–930.
5 Postmus PE, Brambilla E, Chansky K, et al. The IASLC Lung
Cancer Staging Project: proposals for revision of the M descriptors
in the forthcoming (seventh) edition of the TNM classification of
lung cancer. J Thorac Oncol 2007; 2: 686–693.
6 Kassis J, Klominek J, Kohn EC. Tumor microenvironment: what
can effusions teach us? Diagn Cytopathol 2005; 33: 316–319.
7 Zebrowski BK, Yano S, Liu W, et al. Vascular endothelial growth
factor levels and induction of permeability in malignant pleural
effusions. Clin Cancer Res 1999; 5: 3364–3368.
8 Walker-Renard PB, Vaughan LM, Sahn SA. Chemical pleurodesis
for malignant pleural effusions. Ann Intern Med 1994; 120: 56–64.
9 Yeh HH, Lai WW, Chen HH, et al. Autocrine IL-6-induced Stat3
activation contributes to the pathogenesis of lung adenocarcinoma
and malignant pleural effusion. Oncogene 2006; 25: 4300–4309.
10 Chen YM, Tsai CM, Whang-Peng J, et al. Double signal stimulation
was required for full recovery of the autologous tumor-killing effect
of effusion-associated lymphocytes. Chest 2002; 122: 1421–1427.
11 Su LJ, Chang CW, Wu YC, et al. Selection of DDX5 as a novel
internal control for Q-RT-PCR from microarray data using a block
bootstrap re-sampling scheme. BMC Genom 2007; 8: 140.
12 American Joint Committee on Cancer. Manual for Staging of
Cancer. 5th Edn. Philadelphia, JB Lippincott, 1997.
13 Golub TR, Slonim DK, Tamayo P, et al. Molecular classification of
cancer: class discovery and class prediction by gene expression
monitoring. Science 1999; 286: 531–537.
14 Landsittel D, Donohue-Babiak N. Effect of adding fold-change
criteria to significance testing of microarray data. J Statist Comput
Simul 2010; 80: 89–97.
15 Kerstens HM, Robben JC, Poddighe PJ, et al. AgarCyto: a novel
cell-processing method for multiple molecular diagnostic analyses
of the uterine cervix. J Histochem Cytochem 2000; 48: 709–718.
16 Al-Shahrour F, Minguez P, Tarraga J, et al. BABELOMICS: a systems
biology perspective in the functional annotation of genome-scale
experiments. Nucleic acids research 2006; 34: W472–W476.
17 Jimenez-Marin A, Collado-Romero M, Ramirez-Boo M, et al.
Biological pathway analysis by ArrayUnlock and Ingenuity
Pathway Analysis. BMC Proc 2009; 3: Suppl. 4, S6.
18 Barnes DM, Harris WH, Smith P, et al. Immunohistochemical
determination of oestrogen receptor: comparison of different
methods of assessment of staining and correlation with clinical
outcome of breast cancer patients. Brit J Cancer 1996; 74: 1445–1451.
19 Gomez-Fernandez C, Jorda M, Delgado PI, et al. Thyroid
transcription factor 1: a marker for lung adenoarinoma in body
cavity fluids. Cancer 2002; 96: 289–293.
20 Szczepulska-Wojcik E, Langfort R, Roszkowski-Sliz K. [A comparative evaluation of immunohistochemical markers for the differential
diagnosis between malignant mesothelioma, non-small cell carcinoma
EUROPEAN RESPIRATORY JOURNAL
THORACIC ONCOLOGY
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
involving the pleura, and benign reactive mesothelial cell proliferation]. Pneumonol Alergol Pol 2007; 75: 57–69.
Chen YF, Chou CY, Wilkins RJ, et al. Motor protein-dependent
membrane trafficking of KCl cotransporter-4 is important for
cancer cell invasion. Cancer Res 2009; 69: 8585–8593.
Rais B, Comin B, Puigjaner J, et al. Oxythiamine and dehydroepiandrosterone induce a G1 phase cycle arrest in Ehrlich’s tumor
cells through inhibition of the pentose cycle. FEBS Lett 1999; 456:
113–118.
Warburg O. On the origin of cancer cells. Science 1956; 123: 309–314.
Sanchez-Armengol A, Rodriguez-Panadero F. Survival and talc
pleurodesis in metastatic pleural carcinoma, revisited. Report of
125 cases. Chest 1993; 104: 1482–1485.
Duysinx B, Corhay JL, Larock MP, et al. Prognostic value of
metabolic imaging in non-small cell lung cancers with neoplasic
pleural effusion. Nucl Med Commun 2008; 29: 982–986.
Langbein S, Zerilli M, Zur Hausen A, et al. Expression of
transketolase TKTL1 predicts colon and urothelial cancer patient
survival: Warburg effect reinterpreted. Brit J Cancer 2006; 94: 578–585.
Tammali R, Ramana KV, Srivastava SK. Aldose reductase
regulates TNF-alpha-induced PGE2 production in human colon
cancer cells. Cancer Lett 2007; 252: 299–306.
Kroemer G, Pouyssegur J. Tumor cell metabolism: cancer’s
Achilles’ heel. Cancer Cell 2008; 13: 472–482.
Ryan HE, Lo J, Johnson RS. HIF-1 alpha is required for solid
tumor formation and embryonic vascularization. EMBO J 1998; 17:
3005–3015.
Lutz NW, Tome ME, Cozzone PJ. Early changes in glucose and
phospholipid metabolism following apoptosis induction by IFNc/TNF-a in HT-29 cells. FEBS Lett 2003; 544: 123–128.
Stathopoulos GT, Kollintza A, Moschos C, et al. Tumor necrosis
factor-alpha promotes malignant pleural effusion. Cancer Res 2007;
67: 9825–9834.
Hunt TK, Aslam R, Hussain Z, et al. Lactate, with oxygen, incites
angiogenesis. Adv Exp Med Biol 2008; 614: 73–80.
Ojika T, Imaizumi M, Watanabe H, et al. [An immunohistochemical study on three aldolase isozymes in human lung cancer].
Zasshi J 1992; 40: 382–386.
Mori S, Chang JT, Andrechek ER, et al. Anchorage-independent
cell growth signature identifies tumors with metastatic potential.
Oncogene 2009; 28: 2796–2805.
Zhang X, Xu LH, Yu Q. Cell aggregation induces phosphorylation
of PECAM-1 and Pyk2 and promotes tumor cell anchorageindependent growth. Mol Cancer 2010; 9: 7.
Katare R, Andrea C, Emanueli C, et al. Benfotiamine improves
functional recovery of the infarcted heart via activation of prosurvival G6PD/Akt signaling pathway and modulation of
neurohormonal response. J Mol Cell Cardiol 2010; 49: 625–638.
Kwiatkowski DJ, Manning BD. Tuberous sclerosis: a GAP at the
crossroads of multiple signaling pathways. Hum Mol Genet 2005;
14: R251–R258.
Ji H, Ramsey MR, Hayes DN, et al. LKB1 modulates lung cancer
differentiation and metastasis. Nature 2007; 448: 807–810.
Lee DF, Kuo HP, Chen CT, et al. IKK beta suppression of TSC1
links inflammation and tumor angiogenesis via the mTOR pathway. Cell 2007; 130: 440–455.
VOLUME 37 NUMBER 6
1465